MatterSpace Engine
From target properties to viable candidates.
Use these docs when the problem is not generation volume but invalid work. MatterSpace is the Universal Generation Engine for Science and Engineering and a goal-driven inverse generation engine. It turns target properties and hard constraints into candidate sets worth ranking, simulating, or synthesizing.
Overview
What is MatterSpace?
MatterSpace is a goal-driven generation engine for candidates that satisfy domain rules by construction. Use it when conventional generative methods spend most of their budget producing invalid outputs.
The core engine is domain-agnostic. MatterSpace uses one generation workflow to navigate complex search spaces, while the selected category supplies the domain-specific physics, constraints, and objectives.
Lattice is ready now for materials and energy workflows. Vital is available for early technical evaluation in longevity and epigenetic reprogramming workflows.
Core Engine
- Intelligent search space navigation
- Proprietary candidate generation
- Constraint enforcement during generation
- Multi-objective candidate optimization
- Deterministic replay and provenance
Domain Pack
- Domain-specific computational models
- Physical constraints (symmetry, bonds, charges)
- Objective functions and scoring criteria
- Domain-specific generation components
- Validation tiers and acceptance criteria
Public Categories
Lattice and Vital are ready.
Each public category supplies the physics, constraints, objectives, and samplers for a specific field. The core engine remains the same. What changes is the science and the readiness level.
MatterSpace Lattice
Materials Discovery Engine
Crystal structures, alloys, coatings, electrolytes, superconductors, photovoltaics, thermoelectrics, catalysts, magnets, and other functional-material workflows. Ready today.
ReadyMatterSpace Vital
Longevity & Epigenetic Reprogramming Engine
Intervention generation for rejuvenation, epigenetic reprogramming, and delivery-design workflows where safety, reversibility, and biological plausibility must stay inside the search process.
ReadyGeneration Pipeline
From target to archive
Define what you want. MatterSpace selects the right pipeline, keeps constraints inside generation, and returns candidates plus the evidence needed to compare runs later.
Define
Agent or human specifies target properties, constraints, and objectives. MatterSpace auto-selects the right science profile, dynamics parameters, and campaign mode.
Generate
Candidate structures are sampled from the domain-specific compositional and structural search space. Initial configurations respect symmetry and stoichiometry constraints.
Navigate
The engine navigates the search space with an adaptive dynamics controller that selects the right strategy in real time.
Enforce
Physical constraints are enforced during navigation, not after. Bond lengths, coordination numbers, symmetry groups, charge neutrality — validated at every step.
Optimize
Multi-objective optimization across competing properties. Not a single best answer — a diverse set of optimal candidates trading off real-world constraints.
Replay
Every candidate is a typed, provenanced artifact. Full configuration snapshots, dynamics trajectories, constraint satisfaction records, and deterministic replay recipes.
Key Pipeline Properties
Search Space Navigation
The engine generates candidates using constraint-aware search methods, targeting physically stable configurations.
Constraint Enforcement
Physical constraints — bond lengths, coordination numbers, symmetry groups, charge neutrality — are enforced during generation at every step, not applied as post-hoc filters.
Multi-Objective Optimization
The engine maintains a diverse set of optimal candidates. Trade-offs between competing objectives (conductivity vs. stability, hardness vs. ductility) are explored systematically.
Deterministic Replay
Every campaign produces deterministic replay recipes. Configuration snapshots, dynamics trajectories, random seeds, and constraint satisfaction records enable exact reproduction.
Campaign Modes
Choose the campaign mode by the job
Each campaign mode exists for a different design situation: greenfield search, refinement around an anchor, guided rediscovery, or strict benchmark evaluation.
Open Discovery
Use this when there is no anchor candidate and the job is greenfield search under physics constraints. MatterSpace maximizes diversity across the viable frontier.
Prototype Optimization
Use this when you already have a promising anchor candidate and want to refine nearby variants without reopening the whole space.
Guided Rediscovery
Use this when a known class or target helps steer the search and the goal is validation against established science or prior internal results.
Blind Rediscovery Benchmark
Use this for the strictest benchmark path. Targets stay hidden from generation and are revealed only for post-hoc evaluation of whether MatterSpace reached known viable structures.
Early Testing
Programmatic surfaces available for early testing
Programmatic access to MatterSpace is available for early-testing customers. API, SDK, and MCP integration docs are being published as each surface stabilizes.
API Reference
Early TestingComplete OpenAPI specification for the MatterSpace REST API. Campaign management, candidate retrieval, public-category selection, and artifact download endpoints.
Python SDK
Early TestingTyped Python client for MatterSpace. Define campaigns, stream results, evaluate Pareto fronts, and manage artifacts — all with full IDE autocompletion and type safety.
MCP Integration
Early TestingModel Context Protocol server for MatterSpace. AI agents discover and invoke MatterSpace tools automatically — campaign creation, candidate evaluation, and result interpretation.
Ready to reduce invalid search work?
MatterSpace Lattice is ready for materials discovery. MatterSpace Vital is ready for longevity and epigenetic reprogramming workflows.
MatterSpace is patent pending in the United States and other countries. Vareon, Inc.